Path4gmns

v0.8.7 finalizes the changes introduced in v0.8.7a1 as a stable release with the following updates.

1. Calculate and print out relative UE gap as convergency measure.
2. Support the most common length and speed units for link.csv, which include meter (or m), kilometer (or km), mile (or mi), kmh (or kph), and mph.
3. Add more stable handling over the legacy mode 'p' or 'passenger'.
4. Include Public API to docs.
5. Print out distance unit (i.e., mi) in find_shortest_path(), get_agent_node_path(), and get_agent_link_path().

v0.8.7a1 serves as a hot fix to v0.8.6 over the following aspects.

1. add backwards compatibility over the deprecated agent type of p or passenger.
2. enhance demand loading with InvalidRecord when a record of demand volume is empty.
3. fix potential issue in setup_spnetwork(), which requires ascending order of zone ids by adding sorting to get_zones() in class Network.
4. fix the potential issue that bin_index might not start from zero along with the potential zero division issue, when all zones have the same number of nodes in _synthesize_bin_index().
5. change synthesized demand file name to demand.csv in output_synthesized_demand().
6. update all settings.yml in data directory by changing 'p' to 'a' and 'passenger' to 'auto'.
7. enhance exception message when an invalid agent_no as a result of an invalid agent_id is encountered in _get_agent().
9. output origin node and destination node of each agent in agent_paths.csv while remove toll and travel_time. change header 'volume' to 'OD volume' to avoid confusion.
10. enhance the tutorial with elaboration on the legacy way of loading demand and zone information and some caveats.

v0.8.6 fixes crucial bugs related to the column generation module and greatly enhances the parsing functions. A full list of addressed issues can be found [here](https://github.com/users/jdlph/projects/1).

v0.8.6a2 is a **_hotfix_** to v0.8.6a1 over the bug in _output_equity() where only one line was output.

v0.8.6a1 comes as a **_hotfix_** to v0.8.5 over the following two issues.

1. read_nodes() cannot proceed if a record of column "is_boundary" is empty or nonmetric.
2. misleading code snippet in finding shortest paths when link sequence is expected, where the key word "seq_type" should have been placed before argument 'link'.

Other updates include

1. separate _output_equity() from evaluate_accessibility();
2. optimize the setup for bin_index in read_nodes() and polish README on Synthesize Zones and OD Demand.

1. remove legacy namings inherited from dtalite_s.py;
2. remove useless (legacy) members;
3. remove useless imports automatically added by VS Code;
4. change the value of MAX_LABEL_COST to the largest integer (i.e., 2147483647);
5. change the caller of setup_spnetwork() from read_network() to perform_column_generation();
6. introduce class Zone and update the previous implementation where involves zones accordingly;
7. support loading demand for an agent type and a demand period via load_demand();
8. support loading zone information after synthesis;
9. update all settings.yml in data directory to include "use_link_ffs".